Mitsuhiro Oshiki

Boundary Detection Method for Ultrasound Diagnostic Images Using Region-Growing Approach

We have proposed a boundary detection method for ultrasound diagnostic images using the Region-Growing method, which is mainly applicable to the detection of cysts. Region-Growing method, which extracts the region of interest using the images brightness value, is optimized to be applicable to ultrasound diagnostic images. There are some cases in which the boundary is unclear in ultrasound diagnostic images, because the difference in acoustic impedance between two neighboring areas of tissue is small. In such cases, a satisfactory result cannot be obtained only by using the intensity value of the ultrasound diagnostic image. By applying the signal processing to enhance the brightness value transition, we can more easily find the boundary accurately. We adapted this method to be applicable to the diagnostic images of, for example, hyper, hypo, cystic, and mixed echoic lesions, and we confirmed that the boundary of the region of interest is obtained.